Material decomposition for dual-energy propagation-based phase-contrast CT

TL;DR

This paper introduces an iterative, one-step method for material decomposition in dual-energy propagation-based phase-contrast CT, improving accuracy and noise reduction over traditional two-step approaches.

Contribution

It proposes a novel iterative approach that combines phase retrieval, reconstruction, and material decomposition into a single process, enhancing performance.

Findings

Superior accuracy in material decomposition

Reduced noise and artifacts

Effective integration of phase retrieval and reconstruction

Abstract

Material decomposition refers to using the energy dependence of material physical properties to differentiate materials in a sample, which is a very important application in computed tomography(CT). In propagation-based X-ray phase-contrast CT, the phase retrieval and Reconstruction are always independent. Moreover, like in conventional CT, the material decomposition methods in this technique can be classified into two types based on pre-reconstruction and post-reconstruction (two-step). The CT images often suffer from noise and artifacts in those methods because of no feedback and correction from the intensity data. This work investigates an iterative method to obtain material decomposition directly from the intensity data in different energies, which means that we perform phase retrieval, reconstruction and material decomposition in a one step. Fresnel diffraction is applied to forward propagation and CT images interact with this intensity data throughout the iterative process. Experiments results demonstrate that compared with two-step methods, the proposed method is superior in accurate material decomposition and noise reduction.